1
Channel-pore cation selectivity is a major determinant of Bacillus thuringiensis Cry46Ab 1
mosquitocidal activity.
2 3
Tohru Hayakawa*, Midoka Miyazaki, Syoya Harada, Mami Asakura, Toru Ide 4
5
Graduate School of Interdisciplinary Science and Engineering in Health Systems, Okayama 6
University, 3-1-1 Tsushima-naka, Kita-ku, Okayama 700-8530, Japan 7
8
*Corresponding author 9
Tohru Hayakawa, Email address: [email protected] 10
11 12 13
2 Abstract
14
Cry46Ab from Bacillus thuringiensis TK-E6 is a new mosquitocidal toxin with an aerolysin-type 15
architecture, and it is expected to be used as a novel bioinsecticide. Cry46Ab acts as a functional 16
pore-forming toxin, and characteristics of the resulting channel-pores, including ion selectivity, have 17
been analyzed. However, the relationship between channel-pore ion selectivity and insecticidal 18
activity remains to be elucidated. To clarify the effects of charged amino acid residues on the ion 19
permeability of channel-pores and the resulting insecticidal activity, in the present study, we 20
constructed Cry46Ab mutants in which a charged amino acid residue within a putative 21
transmembrane β-hairpin region was replaced with an oppositely charged residue. Bioassays using 22
Culex pipiens mosquito larvae revealed that the mosquitocidal activity was altered by the mutation.
23
A K155E Cry46Ab mutant exhibited toxicity apparently higher than that of wild-type Cry46Ab, but 24
the E159K and E163K mutants exhibited decreased toxicity. Ions selectivity measurements 25
demonstrated that the channel-pores formed by both wild-type and mutant Cry46Abs were cation 26
selective, and their cation preference was also similar. However, the degree of cation selectivity was 27
apparently higher in channel-pores formed by the K155E mutant, and reduced selectivity was 28
observed with the E159K and E163K mutants. Our data suggest that channel-pore cation selectivity 29
is a major determinant of Cry46Ab mosquitocidal activity and that cation selectivity can be 30
controlled via mutagenesis targeting the transmembrane β-hairpin region.
31 32
Key points 33
1. Cry46Ab mutants were constructed by targeting the putative transmembrane β-hairpin region.
34
2. Charged residues within the β-hairpin control the flux of ions through channel-pores.
35
3. Channel-pore cation selectivity is correlated with insecticidal activity.
36 37
Keywords 38
Bacillus thuringiensis TK-E6; Cry46Ab toxin; Culex pipiens mosquito larvae; site-directed 39
mutagenesis; electrophysiologic analysis 40
3 Introduction
41
Cry46Ab is a crystal protein derived from Bacillus thuringiensis strain TK-E6. Cry46Ab has 42
recently renamed as Mpp46Ab1 in new nomenclature (Crickmore et al. 2020). It has been shown 43
that upon activation by proteinase K, Cry46Ab is highly cytotoxic to human leukemic T cells 44
(MOLT-4 and Jurkat), but has virtually no effect on human embryonic kidney cells (HEK293).
45
Cry46Ab was therefore designated parasporin 2Ab, a member of a family of toxins that exhibit 46
preferential cytotoxicity against human cancer cells (Hayakawa et al. 2007). In addition, it was 47
recently reported that Cry46Ab exhibits apparent insecticidal activity against larvae of the mosquito 48
Culex pipiens (Hayakawa et al. 2017). It is noteworthy that co-administration of Cry46Ab with other 49
mosquitocidal Cry toxins, particularly the combination of Cry46Ab and Cry4Aa from B.
50
thuringiensis subsp. israelensis, results in significant synergistic toxicity against C. pipiens larvae 51
(Hayakawa et al. 2017). Cry46Ab is therefore expected to be used not only as a new type of 52
bioinsecticide but also as an agent to enhance the mosquitocidal activity of other Cry toxins.
53
Cry46Ab exhibits significant homology (84% identity) to Cry46Aa (designated 54
parasporin-2Aa) from B. thuringiensis strain A1547 (Hayakawa et al. 2007). Although Cry46Aa is 55
cytotoxic to human leukemic T cells, no insecticidal activity has been reported (Kim et al. 2000). X- 56
ray crystallography analyses revealed that the three-dimensional structure of Cry46Aa is similar to 57
that of aerolysin-type β pore-forming toxins (β-PFTs) (Akiba et al. 2009). Based on its high degree 58
of homology with Cry46Aa, Cry46Ab is thought to be a member of the aerolysin-type β-PFT family 59
(Hayakawa et al. 2007). Previous studies demonstrated that Cry46Ab functions as a PFT, producing 60
cation-selective channel-pores in artificial lipid bilayers (Hayakawa et al. 2017; Sakakibara et al.
61
2019). The reported cation preference of the channel-pores is generally K+ > Na+, K+ > Ca2+, and 62
Ca2+ > Na+ (Sakakibara et al. 2019).
63
Intriguingly, Cry46Ab does not exhibit homology to most other Cry toxins. Indeed, nearly 64
90% of Cry toxins share a characteristic three-domain architecture (domains I, II, and III) and form a 65
large homology group (Schnepf et al. 1998). In general, domain I is located in the N-terminal region 66
and consists of a bundle of seven amphipathic α-helices. The α-helices of domain I are thought to 67
4
form a transmembrane pore, and therefore, these three-domain Cry toxins are classified as α-PFTs.
68
Domain II, which consists of three antiparallel β-sheets, is a putative receptor-binding domain.
69
Domain III, located in the C-terminal region, contains two antiparallel β-sheets that form a β- 70
sandwich fold with a jellyroll topology. Domain III is assumed to be involved in controlling 71
structural integrity and/or receptor binding (Schnepf et al. 1998). Thus, despite the differences in 72
their structures, both aerolysin-type Cry46Ab and three-domain Cry toxins are thought to function 73
via a similar insecticidal mechanism involving pore formation. According to the colloid-osmotic 74
lysis model, pores formed by Cry toxins allow ions and water to pass into target cells, resulting in 75
disruption of the membrane potential, followed by swelling, lysis, and the eventual death of the host 76
cell (Knowles 1994; Knowles and Ellar 1987). On the other hand, the characteristics of the channel- 77
pores formed by Cry toxins have not been investigated in detail. Furthermore, the correlation 78
between channel-pore formation and insecticidal activity is not fully understood.
79
A β-hairpin structure in the middle domain is a characteristic of aerolysin-type β-PFTs.
80
Similar structures have been found in many aerolysin-type β-PFTs, such as aerolysin (Iacovache et 81
al. 2006), staphylococcal α-toxin (Song et al. 1996), enterotoxin from Clostridium perfringens 82
(Kitadokoro et al. 2011), ε-toxin from C. perfringens (Cole et al. 2004), hemolytic lectin from 83
parasitic mushroom Laetiporus sulphureus (Mancheño et al. 2004), leukocidin (Miles et al. 2002), 84
and Cry46Aa (Akiba et al. 2009). According to the pore-formation model of aerolysin, after binding 85
to glycosylphosphatidylinositol-anchored receptor proteins on the target cell membrane, the β- 86
hairpin inserts into the membrane and rearranges into a transmembrane β-barrel (Degiacomi et al.
87
2013; Xu et al. 2014;
Rossjohn
et al. 1998). In general, the β-hairpin is composed of an alternating 88pattern of polar and hydrophobic amino acid residues, suggesting that it is amphipathic. The polar 89
and hydrophobic residues are believed to face the hydrophilic lumen and lipid bilayer of the 90
transmembrane β-barrel, respectively. It has been proposed that the charged amino acid residues 91
within the transmembrane β-hairpin control the flux of ions through the channel-pores (Benz and 92
Popoff 2018). Indeed, the transmembrane β-hairpin of aerolysin contains an excess of positively 93
charged residues (four lysine residues and three glutamic acid residues) and forms anion-selective 94
5
channel-pores (Chakraborty et al. 1990). Similarly, the corresponding region of C. perfringens ε- 95
toxin contains an excess of positively charged residues (one lysine residue and no negatively charged 96
residues) and forms anion-selective channel-pores (Petit et al. 2001). In contrast, the β-hairpin region 97
of C. perfringens enterotoxin contains an excess of negatively charged amino acid residues (no 98
positively charged residues and three glutamic acid residues) and forms cation-selective channel- 99
pores (Kitadokoro et al. 2011).
100
In the present study, we predicted the transmembrane β-hairpin region of Cry46Ab based 101
on sequence alignment analysis with the closely related Cry46Aa, and constructed four Cry46Ab 102
mutants (K155E, K156E, E159K, and E163K) one of the charged amino acid residue in the putative 103
transmembrane region was replaced with an oppositely charged residue. These charged amino acid 104
residues were assumed to line the lumen side of the channel-pores and thus affect the ion 105
permeability of the pores. To investigate the effect of the charged amino acid residues in the 106
transmembrane domain of Cry46Ab on channel-pore ion permeability and clarify the relationship 107
between channel-pore ion permeability and insecticidal activity, the Cry46Ab mutants were 108
subjected to bioassays using Culex pipiens mosquito larvae and ion-selectivity measurements using 109
planar lipid bilayers.
110 111
Materials and methods 112
Construction of the Cry46Ab mutants 113
In the structural model of Cry46Aa, which is most closely related to Cry46Ab, the transmembrane 114
domain is thought to be a β-hairpin region consisting of β8-loop-β9 (Fig. 1a, Akiba et al. 2009). The 115
corresponding region in Cry46Ab spans residues L152 to T168 and contains two positively charged 116
lysine residues (K155 and K156) and two negatively charged glutamic acid residues (E159 and E163) 117
(Fig. 1b).
118
To investigate effect of these charged amino acids on the ion permeability of channel-pores 119
formed by Cry46Ab and on the insecticidal activity resulting from the formation of channel-pores by 120
Cry46Ab, in the present study, we constructed four Cry46Ab substitution mutants (K155E, K156E, 121
6
E159K, and E163K). In these mutants, one charged amino acid was replaced with an oppositely 122
charged amino acid (Fig. 1b). The mutations were introduced via site-directed mutagenesis, as 123
reported previously (Howlader et al. 2009). The expression vector, pGST-Cry46Ab-S1 (Hayakawa et 124
al. 2017) was used as a template. The primers used for mutagenesis are listed in Table 1. Introduction 125
of the desired mutations was confirmed by DNA sequencing.
126 127
Preparation of Cry46Ab toxins 128
Wild-type and mutant Cry46Abs were expressed as glutathione S-transferase (GST) fusions in 129
Escherichia coli BL21 and purified as described previously (Hayakawa et al. 2017). Briefly, E. coli 130
cells were cultured in TB medium containing ampicillin (100 µg/mL) until the OD600 reached 0.5- 131
0.7, and then expression of the GST-Cry46Abs was induced by incubation in 0.1 mM isopropyl-β-D- 132
thiogalactopyranoside at 30°C for 4 h. The E. coli cells were harvested by centrifugation and then 133
disrupted by sonication, and the GST-Cry46Abs were purified using glutathione-Sepharose 4B (GE 134
Healthcare Bio-Sciences AB, Uppsala, Sweden) according to the manufacturer’s instructions. The 135
GST-Cry46Abs were then activated by passage through an immobilized-trypsin column prepared as 136
described previously (Hayakawa et al. 2017). The activated Cry46Abs (polypeptides of 29 kDa) 137
were concentrated using Vivaspin 6 (10-kDa MWCO) centrifugal filter devices (GE Healthcare, 138
Little Chalfont, UK). Protein concentration was estimated using a protein assay kit (Bio-Rad 139
Laboratories, Inc., Hercules, CA) with bovine serum albumin as the standard, and the purified 140
peptides were analyzed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE).
141 142
Measurement of mosquito-larvicidal activity 143
To determine the mosquito-larvicidal activity, purified GST-Cry46Ab wild-type and mutants were 144
administered to Culex pipiens larvae (3rd instar) as described previously (Hayakawa et al. 2017).
145
Mosquito larvae were reared from eggs supplied by the Research and Development Laboratory, 146
Dainihon Jochugiku Co., Ltd. (Osaka, Japan). Mortality was recorded 48 h after toxin addition. The 147
experiments were repeated three times independently, and the average and standard deviation of the 148
7
mortality data were calculated. The 50% lethal dose (LC50) was determined using PROBIT analysis 149
(Finney 1971).
150 151
Ion selectivity measurements 152
Characteristics of the channel-pores formed by Cry46Abs were analyzed using a previously 153
described instrument (Hayakawa et al. 2017; Sakakibara et al. 2019). Briefly, the instrument 154
consisted of two chambers (upper, cis chamber; lower, trans chamber), and both chambers were held 155
at virtual ground, such that the voltage in the solution of the cis chamber was connected to a patch- 156
clamp amplifier by an Ag/AgCl electrode-defined membrane potential. The bottom of the cis 157
chamber was a thin sheet of polyvinyl chloride with a small circular hole (approximately φ200 µm), 158
and a lipid bilayer was constructed by painting asolectin (phospholipids from soybean, Sigma- 159
Aldrich Corp.) solution (40 mg/mL in n-decane) across the small hole. At the same time, liposomes 160
consisting of asolectin were prepared in solution containing 1 M sucrose, as described previously 161
(Sakakibara et al. 2019).
162
To constitute Cry46Ab channel-pores in the lipid bilayer, trypsin-activated Cry46Abs were mixed 163
with liposome solution at a concentration of 25 μg/mL. A previous study suggested that the channel- 164
pores formed by Cry46Ab affect the integrity of lipid bilayer membranes and disrupt liposomes 165
(Sakakibara et al. 2019). The mixture (Cry46Ab and liposome) was added to the solution in the cis 166
chamber to facilitate fusion between the liposomes and the planar lipid bilayer. To analyze anion- 167
cation selectivity of channel-pores formed by Cry46Abs, membrane currents were recorded using a 168
4-fold gradient of KCl across the lipid bilayer (600 mM KCl and 10 mM Tris-HCl [pH 8.0] in the cis 169
chamber, 150 mM KCl and 10 mM Tris-HCl [pH 8.0] in the trans chamber). In addition, to analyze 170
cation preference (K+, Na+, or Ca2+) of channel-pores, different salt solutions (KCl, NaCl, or CaCl2) 171
were used in the cis and trans chambers. KCl and NaCl were used at a concentration of 150 mM, 172
and CaCl2 was used at a concentration of 75 mM to provide Cl− at a concentration equivalent to that 173
in the KCl and NaCl solutions. Data were analyzed using pClamp software (Axon Instruments, 174
Roster City, CA). The current amplitude of the resolvable steps was recorded for each experiment, 175
8
and the resulting data were plotted versus the corresponding applied voltage to generate current- 176
voltage relationships. The zero-current reversal potential (VR) was obtained as the X-intercept of the 177
current-voltage relationship line. The permeability ratio was calculated using the Goldman-Hodgkin- 178
Katz equation.
179 180
Results 181
Preparation of wild-type and mutant Cry46Abs 182
Wild-type and mutant Cry46Abs were expressed as GST fusions in E. coli. SDS-PAGE analysis 183
indicated that the molecular mass of the purified GST-Cry46Abs was approximately 60 kDa, very 184
similar to the expected mass (59.309 kDa) (Fig. 2a). In addition, several protein bands of higher 185
molecular mass suggestive of homodimer formation were observed, particularly in the wild-type 186
GST-Cry46Ab and K155E and K156E mutant samples (Fig. 2a).
187
The wild-type and mutant GST-Cry46Abs were then subjected to treatment using an 188
immobilized-trypsin column. As Cry toxins are activated by trypsin-like proteases in the midgut 189
juice of susceptible insect larvae, this assay serves as a presumptive test of folding fidelity (Almond 190
and Dean 1993). SDS-PAGE analysis revealed that wild-type GST-Cry46Ab was quickly (within 15 191
min) processed into a polypeptide of 29 kDa by this treatment (Fig. 2b). The 29 kDa polypeptide 192
was very similar in size to activated Cry46Ab as reported previously (Hayakawa et al. 2007; 2017) 193
and remained stable for at least 30 min (Fig. 2b). The K155E, E159K, and E163K mutants exhibited 194
a processing pattern very similar to that of wild-type Cry46Ab, suggesting high folding fidelity 195
compared with the wild type (Fig. 2b). However, the K156E mutant was apparently over-processed, 196
such that the amount of 29-kDa polypeptide remaining after 15 min was undetectable by SDS-PAGE 197
(Fig. 2b). This suggested that folding fidelity of overall toxin structure was disrupted by replacement 198
of K156 with E.
199 200
Mosquitocidal activity of wild-type and mutant Cry46Abs 201
The mosquitocidal activity of the wild-type and mutant GST-Cry46Abs was assayed using C. pipiens 202
9
larvae. Purified GST was used as a negative control and exhibited no toxicity at concentrations up to 203
2 µg/mL (data not shown). Wild-type Cry46Ab exhibited toxicity against C. pipiens larvae, with an 204
LC50 value (95% confidence limits) of 0.98 (0.95-1.02) µg/mL (Fig. 3). This LC50 value was very 205
similar to that (1.02 µg/mL) reported previously (Hayakawa et al. 2017).
206
Interestingly, the K155E mutant exhibited toxicity apparently higher than that of wild-type 207
Cry46Ab, with an LC50 value (95% confidence limits) of 0.54 (0.52-0.56) µg/mL (Fig. 3). The 208
remaining mutants, particularly the E163K mutant, exhibited lower toxicity against C. pipiens larvae 209
compared with the wild type. The LC50 values (95% confidence limits) for the K156E, E159K, and 210
E163K mutants were 1.90 (1.80-2.01), 1.53 (1.45-1.63), and 2.74 (2.50-3.06) µg/mL, respectively 211
(Fig. 3). In the case of the K156E mutant, excessive degradation in the midgut juice of C. pipiens 212
larvae was thought to be responsible for the lower toxicity. Therefore, the K156E mutant was not 213
subjected to further analysis. In contrast, the remaining mutants (K155E, E159K, and E163K) 214
exhibited stability upon trypsin treatment comparable to that of wild-type Cry46Ab (Fig. 2B), 215
suggesting that the observed difference in toxicity was due to changes in one or more characteristics 216
of the channel-pores.
217 218
Anion-cation selectivity of channel-pores formed by wild-type Cry46Ab 219
Interestingly, two different current amplitudes were observed in the measurements. One current 220
amplitude was similar to those observed previously (Hayakawa et al. 2017; Sakakibara et al. 2019), 221
characterized as a rapid flickering between open and closed states (Fig. 4a). In previous 222
measurements, activated wild-type Cry46Ab was added directly to the solution in the cis chamber, 223
and the current amplitude of this type was thought to be generated by channel-pores that were 224
directly constituted in the planar lipid bilayer. The current amplitude of the resolvable steps was 225
recoded, pooled for seven independent experiments, and plotted versus the corresponding applied 226
voltage to generate a current-voltage relationship. The current-voltage relationship was a linear, and 227
the channel conductance and VR value were 750 pS and −11.82 mV, respectively (Fig. 4b). The 228
PK/PCl permeability ratio calculated from this VR value was 2.21, demonstrating a higher 229
10
permeability for K+ than Cl−. Formation of cation-selective channel-pores by wild-type Cry46Ab 230
was observed in previous measurements (Sakakibara et al. 2019).
231
The second type of current amplitude was very stable, remaining in the open state for at 232
least several minutes (Fig. 4c). This type of current amplitude was much larger than that described 233
above (Fig. 4a and c), suggesting that multiple channel-pores were formed in the planar lipid bilayer.
234
After the formation of channel-pores in the liposomes, only those liposomes that were destabilized 235
by the formation of multiple channel-pores seemed to fuse with the planar membrane. The current- 236
voltage relationship was a linear, with different conductance levels (ranging from 3.40 to 5.32 nS) in 237
each measurement (Fig. 4d). This suggested that the number of channel-pores formed in the 238
liposomes varied in each measurement. The VR value was −9.38 ± 0.66 mV (n = 7 independent 239
measurements), and the PK/PCl permeability ratio calculated from this VR value was 1.86, 240
demonstrating a higher permeability for K+ than Cl−. Because the VR values were very similar for 241
both types of current amplitudes, these current amplitudes were thought to be generated by the same 242
type of channel-pores. The current amplitudes that stably remained in the open state were subjected 243
to further analysis using channel-pores formed by the Cry46Ab mutants.
244 245
Anion-cation selectivity of channel-pores formed by Cry46Ab mutants 246
Membrane currents through the channel-pores formed by Cry46Ab mutants were recorded as 247
conducted for wild-type Cry46Ab and plotted versus the corresponding applied voltage. The current- 248
voltage relationships for the channel-pores formed by the mutants were linear, with different 249
conductance levels in each measurement (Fig. 5).
250
The VR values obtained with the Cry46Ab mutants varied. The VR value for channel-pores 251
formed by the K155E mutant was −17,06 ± 2.82 mV (n = 5), with a calculated PK/PCl permeability 252
ratio of 3.29 (Fig. 5). This PK/PCl permeability ratio was apparently greater than that of wild-type 253
Cry46Ab, suggesting a much higher permeability for K+ than Cl−. Collectively, these data suggested 254
that replacement of residue K155 with an E residue in the putative transmembrane domain of 255
Cry46Ab increased the negative charge in the channel-pores, resulting in higher permeability for K+ 256
11 than Cl−.
257
In contrast, the VR values for the channel-pores formed by the E159K and E163K mutants 258
were similar, at −6.17 ± 1.58 mV (n = 5) and −5.66 ± 2.30 mV (n = 11), respectively (Fig. 5). The 259
calculated PK/PCl permeability ratios for the E159K and E163K mutants were 1.50 and 1.45, 260
respectively, slightly lower than that of wild-type Cry46Ab and significantly lower than that of the 261
K155E mutant. This suggested that, contrary to the case of the K155E mutant, replacement of 262
residues E159 and E163 with K residue in the putative transmembrane domain increased the positive 263
charge in the channel-pores, resulting in limited permeability of K+. 264
265
Cation preference 266
When the cis chamber contained 150 mM KCl and the trans chamber 150 mM NaCl, the VR values 267
for the channel-pores formed by wild-type Cry46Ab and the K155E, E159K, and E163K mutants 268
were −10.51 ± 0.94 (n = 3), −4.09 ± 1.00 (n = 4), −2.54 ± 1.67 (n = 4), and −5.41 ± 2.12 mV (n = 3), 269
respectively (Fig. 6). The PK/PNa permeability ratios were calculated using the above VR values and 270
corresponding PK/PCl permeability ratios and determined to be 2.07 (wild-type), 1.24 (K155E), 1.19 271
(E159K), and 1.47 (E163K), respectively (Table 2). This indicated that the channel-pores formed by 272
the wild-type and mutant Cry46Abs exhibit a preference for K+ over Na+. In addition, the mutations 273
appeared to reduce the selectivity.
274
Similarly, when the cis chamber contained 75 mM CaCl2 and the trans chamber 150 mM 275
KCl, the VR values for channel-pores formed by wild-type Cry46Ab and the K155E, E159K, and 276
E163K mutants were 2.08 ± 0.84 (n = 3), 2.59 ± 0.63 (n = 3), 4.75 ± 0.62 (n = 3), and 3.79 ± 1.51 277
mV (n = 3), respectively (Fig. 7). The PK/PCa permeability ratios were calculated using the above VR
278
values and corresponding PK/PCl permeability ratios and determined to be 1.18 (wild-type), 1.27 279
(K155E), 1.52 (E159K), and 1.39 (E163K), respectively (Table 2). Thus, the wild-type and mutant 280
Cry46Abs formed channel-pores in which the permeability of K+ was slightly higher than that of 281
Ca2+. 282
When the cis chamber contained 75 mM CaCl2 and the trans chamber 150 mM NaCl, the 283
12
VR values for the channel-pores formed by wild-type Cry46Ab and the K155E, E159K, and E163K 284
mutants were −4.83 ± 0.95 (n = 3), −1.81 ± 1.23 (n = 3), −8.00 ± 1.09 (n = 3), and −1.45 ± 0.22 mV 285
(n = 3), respectively (Fig. 8). The PNa/PCa permeability ratios were calculated using the above VR
286
values and PK/PCl and PK/PNa permeability ratios and determined to be 0.63 (wild-type), 0.88 287
(K155E), 0.51 (E159K), and 0.87 (E163K), respectively (Table 2). These data demonstrated that the 288
wild-type and mutant Cry46Abs formed channel-pores in which the permeability of Ca2+ was 289
slightly higher than that of Na+. 290
Collectively, the above data indicate that there was no difference in the cation preference 291
(K+, Na+,or Ca2+) of the channel-pores between the wild-type and mutant Cry46Abs. Although the 292
PK/PNa, PK/PCa, and PNa/PCa permeability ratios differed for some mutants, the differences were not 293
correlated with the differences in mosquitocidal activity.
294 295
Discussion 296
We previously demonstrated that Cry46Ab toxin forms cation-selective channel-pores in planar lipid 297
bilayer, and the characteristics of these channel-pores have been partially characterized (Hayakawa 298
et al. 2017; Sakakibara et al. 2019). In the mode of action of Cry46Ab, pore formation is thought to 299
be a central component, but the relationship between pore formation and the resulting insecticidal 300
activity remains to be elucidated. Therefore, we constructed substitution mutants (K155E, K156E, 301
E159K, and E163E) targeting the transmembrane β-hairpin region of Cry46Ab and investigated the 302
effects of these mutations on the selectivity of ions passing through the channel-pores and on the 303
resulting mosquitocidal activity. Based on sequence alignment analysis with the closely related 304
Cry46Aa (Fig. 1, Hayakawa et al. 2007), the transmembrane β-hairpin region of Cry46Ab was 305
hypothesized to span amino acid residues L152 to T168. This region contains two positively charged 306
lysine residues (K155 and K156) and two negatively charged glutamic acid residues (E159 and E163), 307
and at least some of these residues are thought to face the hydrophilic lumen of the channel-pores 308
and thereby affect ion permeability.
309
Cry46Ab mutants (except K156E) were successfully expressed in E. coli. A bioassay using 310
13
C. pipiens mosquito larvae demonstrated that the toxicity of the K155E mutant (LC50 = 0.54 μg/mL) 311
was apparently higher than that of wild-type Cry46Ab (LC50 = 0.98 μg/mL). Ion selectivity 312
measurements demonstrated that the permeability ratio PK/PCl of the channel-pores formed by the 313
K155E mutant (PK/PCl = 3.29) was apparently higher than that of wild-type Cry46Ab (PK/PCl = 1.86) 314
(Table 2). This suggests that, as expected for the K155E mutant, replacement of a positively charged 315
lysine (K155) with a negatively charged glutamic acid (E) rendered the environment of the lumen 316
more anionic, resulting in the channel-pores becoming more cation selective. In contrast, the toxicity 317
of the E159K (LC50 = 1.53 μg/mL) and E163K (LC50 = 2.74 μg/mL) mutants was apparently lower 318
than that of wild-type Cry46Ab, and their PK/PCl values (1.50 for E159K; 1.45 for E163K) were 319
slightly lower than that of wild-type Cry46Ab (Fig. 5; Table 2). This suggests that replacement of 320
negatively charged glutamic acid residues (E159 and E163) with positively charged lysine residues (K) 321
rendered the environment of the lumen more cationic, resulting in the channel-pores becoming less 322
cation selective. Changing the PK/PCl permeability ratio thus apparently affects mosquitocidal 323
activity, as increasing the selectivity of Cry46Ab resulted in higher toxicity against mosquito larvae.
324
The formation of highly cation-selective channel-pores may enhance the influx of cations and water 325
into the larval cell, thus disrupting the membrane potential and inducing swelling, lysis, and the 326
eventual death of the host cell. Nevertheless, this notion still needs to be elucidated. Further studies 327
using mutants combined either two of mutation K155E, E159K and E163E would be of great 328
interest. It would also be of interest to investigate mutants in which charged residues K155, E159 329
and E163 are replaced with other type of residue such as non-charged residues, and mutants in which 330
a polar residues other than K155, E159 and E163 are replaced.
331
It has been demonstrated that co-administration of Cry46Ab with three-domain Cry toxins, 332
especially the combination of Cry46Ab and Cry4Aa, results in significant synergistic toxicity against 333
C. pipiens larvae (Hayakawa et al. 2017). It is believed that co-administration of multiple toxins 334
exhibiting different modes of action prevents the onset of resistance in insects. Synergistic toxicity is 335
observed when multiple toxins exhibiting different modes of action are co-administrated, suggesting 336
differences in the mode of action of Cry46Ab and three-domain Cry toxins. However, Cry46Ab has 337
14
been demonstrated to function as a PFT, and pore formation has also been demonstrated with several 338
three-domain Cry toxins. According to the umbrella model, helices α4 and α5 of domain I insert into 339
the membrane to form pores, while the remaining helices spread along the outer membrane surface 340
via a conformational change (Gazit et al. 1998). As such, Cry46Ab and three-domain Cry toxins are 341
thought to share a similar insecticidal mechanism based on pore formation. On the other hand, 342
binding receptors seems to be different between aerolysin-type Cry toxins and three-domain Cry 343
toxins (Xu et al. 2014), and the difference may proceed synergistic toxicity. However, the 344
determinants that facilitate synergistic toxicity involving Cry46Ab and three-domain Cry toxins 345
remain to be elucidated.
346
Interestingly, channel-pores formed by wild-type Cry46Ab exhibit a K+ > Na+, K+ > Ca2+, 347
and Ca2+ > Na+ cation preference (Sakakibara et al. 2019). In the present study, this observation was 348
confirmed using a modified ion selectivity measurement procedure (Figs. 6-8). There was no 349
significant difference in channel pore cation preference between the wild-type and mutant 350
Cry46Abs. Although some of the mutations affected the PK/PNa, PK/PCa, and PNa/PCa values, the 351
changes were not correlated with a change in mosquitocidal activity (Table 2). In contrast, it is 352
widely accepted that the influx of ions into cells causes not only osmotic shock but also apoptosis, 353
suggesting that differences in the cation preference of channel-pores may result in different effects 354
on insecticidal activity. The characteristics of channel-pores have not been investigated in detail for 355
three-domain Cry toxins. It would thus be of interest to investigate the cation preference of channel- 356
pores formed by three-domain Cry toxins, especially Cry toxins that exhibit synergistic toxicity with 357
Cry46Ab.
358 359
Contributions 360
TH and TI conceived and designed research. TH and SH constructed mutants and analyzed their 361
biological activity. TH, MM, MA and TI contributed electrophysiologic experiments. TH, MM and 362
TI analyzed data. TH wrote the manuscript. All authors read and approved the manuscript.
363 364
15 Acknowledgements
365
Eggs of C. pipiens were kindly supplied by the Research and Development Laboratory at Dainihon 366
Jochugiku Co., Ltd., Osaka, Japan.
367 368
Declarations 369
Funding information: The present work was supported in part by a research grant from the 370
OSHIMO foundation (2019) and a JSPS KAKENHI grant (number JP18K05675).
371
Conflicts of interest: All authors declare that they have no conflicts of interest.
372
Ethical approval: This article does not describe any studies with human participants or animals 373
performed by any of the authors.
374
Consent to participate: All authors approved participation.
375
Consent for publication: All authors approved publishing of this article.
376
Availability of data and material: Not applicable 377
Code availability: Not applicable 378
379
References 380
Akiba T, Abe Y, Kitada S, Kusaka Y, Ito A, Ichimatsu T, Katayama H, Akao T, Higuchi K, Mizuki E, 381
Ohba M, Kanai R, Harata K (2009) Crystal structure of the parasporin-2 Bacillus thuringiensis 382
toxin that recognizes cancer cells. J Mol Biol 386:121-133. doi: 10.1016/j.jmb.2008.12.002 383
Almond BD, Dean DH (1993) Structural stability of Bacillus thuringiensis delta-endotoxin 384
homolog-scanning mutants determined by susceptibility to proteases. Appl Environ Microbiol 385
59:2442-2448.
386
Benz R, Popoff MR (2018) Clostridium perfringens enterotoxin: The toxin forms highly cation- 387
selective channels in lipid bilayers. Toxins (Basel) 10:E341. doi: 10.3390/toxins10090341 388
Chakraborty T, Schmid A, Notermans S, Benz R (1990) Aerolysin of Aeromonas sobria: Evidence 389
for formation of ion-permeable channels and comparison with alpha-toxin of Staphylococcus 390
aureus. Infect Immun 58:2127-2132.
391
16
Cole AR, Gibert M, Popoff M, Moss DS, Titball RW, Basak AK (2004) Clostridium perfringens 392
epsilon-toxin shows structural similarity to the pore-forming toxin aerolysin. Nat Struct Mol 393
Biol 11:797-798. doi: 10.1038/nsmb804 394
Crickmore N, Berry C, Panneerselvam S, Mishra R, Connor TR, Bonning BC (2020) A structure- 395
based nomenclature for Bacillus thuringiensis and other bacteria-derived pesticidal proteins. J 396
Invertebr Pathol. doi: 10.1016/j.jip.2020.107438 397
Degiacomi MT, Lacovache I, Pernot L, Chami M, Kudryashev M, Stahlberg H, van der Goot FG, 398
dal Peraro M (2013) Molecular assembly of the aerolysin pore reveals a swirling membrane- 399
insertion mechanism. Nat Chem Biol 9:623–629. doi: 10.1038/nchembio.1312 400
Finney DJ (1971) Probit analysis. 3rd ed. Cambridge Univ. Press, London, United 401
Kingdom.
402
Gazit E, La Rocca P, Sansom MS, Shai Y (1998) The structure and organization within the 403
membrane of the helices composing the pore-forming domain of Bacillus thuringiensis delta- 404
endotoxin are consistent with an "umbrella-like" structure of the pore. Proc Natl Acad Sci U S 405
A 95:12289-12294. doi: 10.1073/pnas.95.21.12289 406
Hayakawa T, Kanagawa R, Kotani Y, Yamagiwa M, Kimura M, Yamane Y, Takebe S, Sakai H (2007) 407
Parasporin-2Ab, a newly isolated cytotoxic crystal protein from Bacillus thuringiensis. Curr 408
Microbiol 55:278-283. doi: 10.1007/s00284-013-0301-1 409
Hayakawa T, Sakakibara A, Ueda S, Azuma Y, Ide T, Takebe S (2017) Cry46Ab from Bacillus 410
thuringiensis TK-E6 is a new mosquitocidal toxin with aerolysin-type architecture. Insect 411
Biochem Mol Biol 87:100-106. doi: 10.1016/j.ibmb.2017.06.015 412
Howlader MTH, Kagawa Y, Sakai H, Hayakawa T (2009) Biological properties of loop-replaced 413
mutants of Bacillus thuringiensis mosquitocidal Cry4Aa. J Biosci Bioeng 108:179-283. doi:
414
10.1016/j.jbiosc.2009.03.016 415
Iacovache I, Paumard P, Scheib H, Lesieur C, Sakai N, Matile S, Parker MW, van der Goot FG 416
(2006) A rivet model for channel formation by aerolysin-like pore-forming toxins. EMBO J 417
25:457-466. doi: 10.1038/sj.emboj.7600959 418
17
Kim HS, Yamashita S, Akao T, Saitoh H, Higuchi K, Park YS, Mizuki E, Ohba M (2000) In vitro 419
cytotoxicity of non-cyt inclusion proteins of a Bacillus thuringiensis isolate against human 420
cells, including cancer cells. J Appl Microbiol 89:16-23. doi: 10.1046/j.1365- 421
2672.2000.01087.x 422
Kitadokoro K, Nishimura K, Kamitani S, Fukui-Miyazaki A, Toshima H, Abe H, Kamata Y, Sugita- 423
Konishi Y, Yamamoto S, Karatani H, Horiguchi Y (2011) Crystal structure of Clostridium 424
perfringens enterotoxin displays features of beta-pore-forming toxins. J Biol Chem 286:19549- 425
19555. doi: 10.1074/jbc.M111.228478 426
Knowles BH (1994) Mechanism of action of Bacillus thuringiensis insecticidal δ-endotoxins. Adv 427
Insect Physiol 24:275-308.
428
Knowles BH, Ellar DJ (1987) Colloid-osmotic lysis is a general feature of the mechanism of action 429
of Bacillus thuringiensis δ-endotoxins with different insect specificity. Biochim Biophys Acta 430
924:507–518.
431
de Maagd RA, Bravo A, Berry C, Crickmore N, Schnepf HE (2003) Structure, diversity, and 432
evolution of protein toxins from spore-forming entomopathogenic bacteria. Annu Rev Genet 433
37:409-433.
434
Mancheño JM, Tateno H, Goldstein IJ, Hermoso JA (2004) Crystallization and preliminary 435
crystallographic analysis of a novel haemolytic lectin from the mushroom Laetiporus 436
sulphureus. Acta Crystallogr D Biol Crystallogr 60:1139-1141. doi:
437
10.1107/S0907444904007991 438
Miles G, Movileanu L, Bayley H (2002) Subunit composition of a bicomponent toxin:
439
staphylococcal leukocidin forms an octameric transmembrane pore. Protein Sci 11:894-902.
440
doi: 10.1110/ps.4360102 441
Parker MW, Buckley JT, Postma JP, Tucker AD, Leonard K, Pattus F, Tsernoglou D (1994) Structure 442
of Aeromonas toxin proaerolysin in its water-soluble and membrane-channel states. Nature 443
367:292-295. doi: 10.1038/367292a0 444
Petit L, Maier E, Gibert M, Popoff MR, Benz R (2001) Clostridium perfringens epsilon toxin 445
18
induces a rapid change of cell membrane permeability to ions and forms channels in artificial 446
lipid bilayers. J Biol Chem 276:15736-15740. doi: 10.1074/jbc.M010412200 447
Rossjohn J, Feil SC, McKinstry WJ, Tsernoglou D, van der Goot G, Buckley JT, Parker MW (1998) 448
Aerolysin-a paradigm for membrane insertion of beta-sheet protein toxins? J Struct Biol 449
121:92–100. doi: 10.1006/jsbi.1997.3947 450
Sakakibara A, Takebe S, Ide T, Hayakawa T (2019) Characterization of the channel-pores formed by 451
Bacillus thuringiensis Cry46Ab toxin in planar lipid bilayers. Appl Entomol Zool 54:389-398.
452
doi: 10.1007/s13355-019-00635-z 453
Schnepf E, Crickmore N, Van Rie J, Lereclus D, Baum J, Feitelson J, Zeigler DR, Dean DH (1998) 454
Bacillus thuringiensis and its pesticidal crystal proteins. Microbiol Mol Biol Rev 62:775-806.
455
Song L, Hobaugh MR, Shustak C, Cheley S, Bayley H, Gouaux JE (1996) Structure of 456
staphylococcal alpha-hemolysin, a heptameric transmembrane pore. Science 274:1859-1866.
457
doi: 10.1126/science.274.5294.1859 458
Xu C, Wang BC, Yu Z, Sun M (2014) Structural insights into Bacillus thuringiensis Cry, Cyt and 459
parasporin toxins. Toxins (Basel) 6:2732-2770. doi: 10.3390/toxins6092732 460
461
19 462
Table 1. Nucleotide sequences of the primers used for site-directed mutagenesis.
463
Mutants Primers Primer sequence (5' → 3')
K155E 46Ab-155E-f GAAAAAGTCTTTGAAATTGGT 46Ab-155-156r AATCGACAGTTTAGTGGTAAT
K156E 46Ab156E-f AAAGAAGTCTTTGAAATTGGT 46Ab-155-156r
E159K 46Ab159K-f AAAATTGGTGGCGAAGTTTCG 46Ab159-163r AAAGACTTTTTTAATCGACAG
E163K 46Ab163K-f GAAATTGGTGGCAAAGTTTCG 46Ab159-163r
464 465
20 466
Table 2. Relationship between mosquitocidal activity and ion permeability of channel-pores formed 467
by wild-type and mutant Cry46Abs.
468
Cry46Ab
Mosquitocidal activity Ion permeability ratios
LC50 (μg/mL) 95% confidence interval PK/PCl PK/PNa PK/PCa PNa/PCa
WT 0.98 0.95 - 1.02 1.86 2.07 1.18 0.63
K155E 0.54 0.52 - 0.56 3.29 1.24 1.27 0.88
E159K 1.53 1.45 - 1.63 1.50 1.19 1.09 0.51
E163K 2.74 2.50 - 3.06 1.45 1.47 1.39 0.87
469 470
21 Figure legends
471 472
Fig. 1 Putative transmembrane β-hairpin region of Cry46Ab (A) Structural neighbors of Cry46Ab.
473
Ribbon diagrams are drawn from PDB data under codes 1PRE for aerolysin (Parker et al. 1994), 474
1UYJ for epsilon toxin (Cole et al. 2004), and 2ZTB for Cry46Aa (Akiba et al. 2009). All images for 475
the molecular structure is prepared with PyMOL. Putative transmembrane β-hairpin regions are 476
indicated by black color. (B) Comparison of the putative transmembrane β-hairpin regions of 477
different aerolysin-type β-PFTs. The alignment of aerolysin and epsilon toxin is adapted from 478
Iacovache et al. (2006). Cry46Aa (Kim et al. 2000) and Cry46Ab (Hayakawa et al. 2007) share 479
identical amino acid sequence in the putative transmembrane β-hairpin region. Amino acid residue 480
number is shown at left.
481 482
Fig. 2 Recombinant wild-type and mutant Cry46Abs. (A) Wild-type and mutant GST-Cry46Abs 483
were purified using glutathione beads and analyzed by 10% SDS-PAGE. One microgram of purified 484
protein was applied. (B) Wild-type and mutant Cry46Abs were treated with an immobilized-trypsin 485
column and analyzed by 15% SDS-PAGE.
486 487
Fig. 3 Mosquitocidal activity of wild-type and mutant GST-Cry46Abs. Filled circles, wild-type GST- 488
Cry46Ab; open circles, mutant GST-Cry46Abs. The experiments were repeated independently more 489
than three times. Average (standard deviation) mortality rates observed at 48 h after administration 490
are shown.The LC50 values (95% confidence limits) were determined using PROBIT analysis 491
(Finney, 1971).
492 493
Fig. 4 Anion-cation selectivity of channel-pores formed by wild-type Cry46Ab. Membrane currents 494
though the channel-pores formed by wild-type Cry46Ab were recorded with a 4-fold gradient of KCl 495
across the lipid bilayer. (A) Representative segments of membrane current flickering between open 496
and closed states. Current levels corresponding to the open state of the channel-pores are indicated 497
22
by a dashed line. (B) Current-voltage relationship of membrane current flickering between open and 498
closed states. The zero-current reversal potential (VR) was calculated from the equation of the fitted 499
line. (C) Representative segments of membrane current remain in the open state for an extended 500
time. (D) The current-voltage relationship of membrane current remains in the open state for an 501
extended time. The experiment was repeated 7 times independently, and the average (standard 502
deviation) VR was determined using each fitted line.
503 504
Fig. 5 Anion-cation selectivity of channel-pores formed by Cry46Ab mutants. Membrane currents 505
though channel-pores formed by the Cry46Ab mutants were recorded with a 4-fold gradient of KCl 506
across the lipid bilayer. The experiment was independently repeated 5 times for the K155E and 507
E159K mutants and 11 times for the E163K mutant. The average (standard deviation) VR was 508
determined using each fitted line.
509 510
Fig. 6 Cation selectivity (K+ vs. Na+) of channel-pores formed by Cry46Abs. Membrane currents 511
though the channel-pores formed by Cry46Abs were recorded under asymmetric buffer conditions 512
(150 mM KCl in the cis chamber, 150 mM NaCl in the trans chamber) across the lipid bilayer. The 513
experiment was independently repeated 3 times, and the average (standard deviation) VR was 514
determined using each fitted line.
515 516
Fig. 7 Cation selectivity (Ca2+ vs. K+) of channel-pores formed by Cry46Abs. Membrane currents 517
though channel-pores formed by the Cry46Abs were recorded under asymmetric buffer conditions 518
(75 mM CaCl2 in the cis chamber, 150 mM KCl in the trans chamber) across the lipid bilayer. The 519
experiment was independently repeated 3 times, and the average (standard deviation) VR was 520
determined using each fitted line.
521 522
Fig. 8 Cation selectivity (Ca2+ vs. Na+) of channel-pores formed by Cry46Abs. Membrane currents 523
though channel-pores formed by the Cry46Abs were recorded under asymmetric buffer conditions 524
23
(75 mM CaCl2 in the cis chamber, 150 mM NaCl in the trans chamber) across the lipid bilayer. The 525
experiment was independently repeated 3 times, and the average (standard deviation) VR was 526
determined using each fitted line.
527
